Establishment of a well-characterized SARS-CoV-2 lentiviral pseudovirus neutralization assay using 293T cells with stable expression of ACE2 and TMPRSS2

Pseudoviruses are useful surrogates for highly pathogenic viruses because of their safety, genetic stability, and scalability for screening assays. Many different pseudovirus platforms exist, each with different advantages and limitations. Here we report our efforts to optimize and characterize an HIV-based lentiviral pseudovirus assay for screening neutralizing antibodies for SARS-CoV-2 using a stable 293T cell line expressing human angiotensin converting enzyme 2 (ACE2) and transmembrane serine protease 2 (TMPRSS2). We assessed different target cells, established conditions that generate readouts over at least a two-log range, and confirmed consistent neutralization titers over a range of pseudovirus input. Using reference sera and plasma panels, we evaluated assay precision and showed that our neutralization titers correlate well with results reported in other assays. Overall, our lentiviral assay is relatively simple, scalable, and suitable for a variety of SARS-CoV-2 entry and neutralization screening assays.


Introduction
In December 2019 a cluster of atypical pneumonia cases appeared in Wuhan, China. The etiological agent was later identified as severe acute respiratory syndrome coronavirus 2 (SARS--CoV-2), the causative agent of coronavirus disease 2019 (COVID-19) [1][2][3][4]. In the past year, SARS-CoV-2 spread as a global pandemic with more than 75 million cases and 1.6 million deaths (Source: Johns Hopkins Coronavirus Resource Center; https://coronavirus.jhu.edu/). A key priority in fighting the ongoing pandemic involves measuring immune responses to the spike (S) glycoprotein of SARS-CoV-2, a critical target for developing preventive vaccines [5] and antibody (Ab)-based therapeutics for COVID-19 patients [6,7], including therapeutic monoclonal antibodies (mAbs) and convalescent plasma therapy [8][9][10][11][12][13][14][15][16]. Assessments of serological responses to the S glycoprotein typically include virus microneutralization (MN) assays or enzyme-linked immunosorbent assay (ELISA), and ELISA variants, such as lateral flow assay (LFA), chemiluminescence immunoassay (CLIA), and electrochemiluminescence immunoassay (ECLIA) [17][18][19]. Replicating, wild-type (WT) SARS-CoV-2 MN assays remain the gold standard, but they are labor intensive due to the need for high biosafety level containment (BSL-3) handling by trained personnel and challenges for high throughput [20]. On the other hand, ELISA formats are safe and high throughput, but they do not always measure titers that strongly correlate with neutralization titers measured in the WT MN assay [10,18,[21][22][23]. Neutralizing Abs are thought to be an important component of protection. Some Abs that bind to S in ELISAs do not neutralize virus because they bind to S epitopes that do not interfere with receptor binding or fusion steps needed for virus entry [24][25][26].
The trimeric S glycoprotein mediates virus entry by binding to the ACE2 receptor on target cells and catalyzing fusion between viral and target cell membranes. Proteolytic processing of S is required for its fusion competence. The multi-basic furin-like cleavage site (RRAR � SV) allows S to be efficiently cleaved into the S1 subunit that contains the receptor binding domain (RBD) and the S2 subunit that contains domains needed for fusion [27][28][29]. Efficient entry into the target cells additionally requires S protein priming at the S2' site by cellular proteases, such as TMPRSS2 or cathepsins B and L (Cat B or L) [28]. Depending upon the cell type, cellular proteases promote entry at the cell surface (e.g., TMPRSS2 in lung epithelium and TMPRSS4 in gut enterocytes) or in endosomes (e.g., Cat L) [28,30]. Small molecules or other inhibitors that target the S protein fusion function or cellular proteases needed for S2' priming prevent the fusion step of entry [28]. Neutralizing Abs directed against the top of the RBD typically compete with virus binding to ACE2, while those directed against the side surfaces of the RBD often do not efficiently compete with ACE2 binding and may therefore show less potent neutralization [17].
Pseudoviruses bearing viral envelope proteins provide safe surrogates for highly pathogenic viruses in MN assays. Several groups have generated SARS-CoV-2 pseudoviruses with glycoprotein defective murine leukemia virus (MLV)-, human immunodeficiency virus (HIV)-, and vesicular stomatitis virus (VSV)-based systems and used them in neutralization assays based on fluorescence (monomeric Neon Green or mNG) or enzymatic activity (nano-, gaussia-, and firefly luciferases) read outs in a variety of target cell types [20,24,[31][32][33][34][35][36][37][38][39][40][41][42][43][44]. In the present study, we describe our optimized conditions for an HIV-based lentiviral SARS CoV-2 pseudovirus neutralization assay. To resemble respiratory cells with TMPRSS2 and facilitate assay procedures, we established a stable 293T cell line expressing both ACE2 and TMPRSS2. We present our detailed methodology and the performance characteristics of the assay, which should be suitable for many quantitative, high-throughput virus neutralization and entry screens that can be easily performed in a BSL-2 laboratory.

Antibodies and sera
Mouse mAb 10G6H5 against SARS-COV2 S protein was purchased from GenScript (Piscataway, NJ). Rabbit antisera against the S1 subunit, the receptor binding domain (RBD), and the S2 subunit of SARS-COV2 S protein [49]

Pseudovirus production and neutralization
Pseudoviruses bearing the S glycoprotein and carrying a firefly luciferase (FLuc) reporter gene were produced in 293T cells. Briefly, 5μg of pCMVΔR8.2, 5μg of pHR'CMVLuc and 0.5μg of S or its mutants (codon optimized) expression plasmids with or without 2μg of the TMPRSS2 expression plasmid were co-transfected in 293T cells. Pseudovirus supernatants were collected approximately 48 h post-transfection, filtered through a 0.45μm low protein binding filter, and used immediately or stored at -80˚C. Pseudovirus titers were measured by infecting different cells for 48 h prior to measuring luciferase activity (luciferase assay reagent, Promega, Madison, WI), as described previously [51]. Pseudovirus titers were expressed as relative luminescence unit per milliliter of pseudovirus supernatants (RLU/ml).
Neutralization assays were performed on 293T cells transiently transfected or transduced with ACE2 and TMPRSS2 genes for stable expression. Briefly, pseudoviruses with titers of approximately 10 6 RLU/ml of luciferase activity were incubated with antibodies or sera for one hour at 37˚C. Pseudovirus and antibody mixtures (100 μl) were then inoculated onto 96-well plates that were seeded with 3.0 x 10 4 cells/well one day prior to infection. Pseudovirus infectivity was scored 48 h later for luciferase activity. The antibody dilution or mAb concentration causing a 50% and 80% reduction of RLU compared to control (ID 50 and ID 80 or IC 50 and IC 80 , respectively) were reported as the neutralizing antibody titers. Titers were calculated using a nonlinear regression curve fit (GraphPad Prism software Inc., La Jolla, CA). The mean 50% and 80% reduction of RLU compared to control from at least two independent experiments was reported as the final titer. For experiments involving camostat mesylate (0.03-500 μM) and chloroquine (0.39-25 μM), target cells were treated with each inhibitor for two hours before pseudovirus infection in the presence of respective inhibitor.
To generate stable 293T-ACE2.TMPRSS2 s cells, VSV-G-pseudotyped lentiviruses carrying the human TMPRSS2 gene were generated by co-transfecting 293T cells with the pHAGE2-E-F1aInt-TMPRSS2-IRES-mCherry, pCMVΔ8.2, and pCMV-VSV-G plasmids. Packaged lentivirus was used to transduce 293T-ACE2 cells in the presence of 10μg/mL polybrene, and the resulting bulk transduced population was single-cell sorted into clear bottomed 96 well plates by flow cytometry that was based on intermediate or high mCherry positivity on a BD FAC-SAria II Cell Sorter. Once single cell clones reached confluence, they were screened for mCherry/TMPRSS2 expression via EVOS Floid cell imaging station (Thermo Fisher, Waltham, MA), and several clones with visible mCherry expression were expanded. For verifying mCherry expression via flow cytometry, cells were harvested with enzyme-free cell dissociation buffer (Thermo Fisher, Waltham, MA), washed, and resuspended in FACS buffer. One clone that displayed intermediate levels of mCherry expression and maximum pseudovirus infectivity titer was selected and referred to as 293T-ACE2.TMPRSS2 s . Up to the present, this clone has supported high-level infectivity of SARS-CoV-2 pseudoviruses through 20 passages. This cell clone is deposited in BEI resources under the item number NRS-55293.

SARS-CoV-2 mNG infection and confocal microscopy
Vero E6 cells, 293T-hACE2 s , and 293T-ACE2.TMPRSS2 s cells were seeded on poly-L-lysinecoated coverslips one day prior to infection. Infection with replicating SARS-CoV-2-mNG (MOI:0.1) was carried out in medium containing 2% FBS for one hour at 37˚C, prior to washing the cells twice with PBS and then maintaining in culture, described above. SARS-CoV-2-mNG expressing mNeon Green (mNG) in place of ORF7 was described previously [52]. At 24 h post infection (p.i.) SARS-CoV-2-mNG-infected coverslips were fixed with 4% paraformaldehyde in PBS at room temperature for 20 min followed by PBS washes. SARS--CoV-2-mNG infection and fixing procedures were performed in a BSL-3 laboratory at the US Food and Drug Administration. Coverslips were counterstained with Hoechst 33258 dye (Thermo Scientific) and mounted on microscope slides with Fluoromount-G (SouthernBiotec). Confocal microscopy was performed by using SP8 DMI6000 confocal microscope (Leica Microsystems Inc, Germany) equipped with 25x water immersion objective lens and 405, 488 and 594 laser lines for Hoechst, mNG and mCherry signal, respectively.

Immunoblot analysis
Pseudoviruses were resolved by SDS-PAGE and detected by Western blot using a mouse mAb (183-H12-5C) against HIV-1 p24 Gag and rabbit antisera against the S1 subunit and the S2 subunit of SARS-COV-2 S protein.

Statistics analysis
To evaluate assay precision, six NIBSC plasma standards, 15 focused concordance samples and 21 SNACS samples were tested by three operators. Two operators ran four independent experiments (two independent experiments per operator), and a third operator ran one experiment. Titers were calculated from curves using eight dilutions. Intermediate precision, expressed as the percent coefficient of variation (%CV), was assessed separately for ID 50 and ID 80 titers. Sample dilutions with observed titers of less than 1:40 were considered as negative for antibodies to SARS-CoV-2 and were imputed a value of 1:20. An exploratory analysis was additionally performed by excluding titers of less than 1:40. Samples with more than 50% of less than 1:40 were excluded from all analyses. The total %CV accounts for both inter-operator and interassay variability and were estimated as follows based on a linear mixed model of the natural log-transformed titers with sample as a fixed effect and operator as a random effect: %CV ¼ ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi eŝ 2 OP þŝ 2 whereŝ 2 OP andŝ 2 IS were the estimated inter-operator and inter-assay (within-operator) variance components from the model, respectively. SAS version 9.4 was used to perform the linear mixed model analysis.
To evaluate accuracy, since the true titers of test samples are not known, the Spearman correlation coefficient between the reported titers and the "observed" titers was estimated with GraphPad Prism software.
A prior report demonstrated priming of SARS-CoV-2 S in VSV-based pseudoviruses by TMPRSS2-related proteases TMPRSS11D, 11E, 11F, and 13 in Calu-3 target cells. Of the TMPRSS2-related proteases, TMPRSS13 was as efficient as TMPRSS2, while the others were less efficient in promoting S priming and thus infectivity [41]. In the present study, we investigated the ability of TMPRSS11D (also known as human airway trypsin) in target 293T cells transiently expressing ACE2 and TMPRSS11D (293T-ACE2.TMPRSS11D t ). Consistent with the prior report, TMPRSS11D was less efficient in S priming, as reflected by an 11-fold lower infectivity in 293T-ACE2.TMPRSS11D t cells compared to 293T-ACE2.TMPRSS2 t cells (Fig 1A).
To facilitate SARS-CoV-2 pseudovirus neutralization assays and partly mimic natural SARS-CoV-2 target cells that express TMPRSS2, we established a stable cell line expressing both ACE2 and TMPRSS2 (293T-ACE2.TMPRSS2 S ) by transducing the 293T-ACE2 cells with a lentivirus encoding TMPRSS2 and mCherry as a bicistrionic transcript. The ACE2.TMPRSS2 S cells conferred infectivity approximately 1700-fold above background, confirming the contribution of TMPRSS2 protease activity for priming the S protein for fusion competence. Our findings agree with a recent study reporting that TMPRSS2 enhances SARS-CoV-2 pseudovirus entry and that S activation by TMPRSS2 correlates with the presence of the multi-basic furin cleavage site [55]. Because the ACE2.TMPRSS2 S cells facilitated greater levels of infectivity and provided > 10 8 RLU/ml infectivity (>3 log range) for resolving titers, we focused on qualifying the 293T-ACE2.TMPRSS2 s cells for our future studies. We also confirmed S protein incorporation into pseudoviruses and proteolytic processing of the S protein to generate S1 and S2 subunits that migrate at 130 kDa and 75 kDa, respectively (Fig 1B and Supporting Information).

Optimization of SARS-CoV-2 pseudovirus infectivity
We investigated several conditions for optimizing SARS-CoV-2 pseudovirus production in 293T cells. When comparing S priming by TMPRSS2 during pseudovirus production to priming during entry into target cells, we found that co-expressing TMPRSS2 during pseudovirus production reduced pseudovirus infectivity, possibly due to TMPRSS2-induced premature activation of S that promotes conformational changes to fusion-incompetent or post-fusion structures (Fig 1C). This finding is consistent with a previous report suggesting the importance of tight regulation of protease cleavage of the S protein for preserving SARS-CoV-2 infectivity [56].
We also investigated variant S proteins to further optimize pseudovirus production. We generated S genes with the D614G mutation or a C-terminal cytoplasmic tail truncation of 19 amino acids (TR19) that were previously reported to yield higher infectivity titers compared to full length WT S glycoprotein [34,38,39,42,57]. The D614G change represents a natural mutation outside RBD that became dominant in the circulating strains [39]. In 293T-ACE2 s cells, the D614G mutation was reported to confer 1-log or >1/2-log higher infectivity to VSV-based or lentivirus-based S pseudoviruses, respectively [38,39,58]. Furthermore, viruses with the G614 S were associated with higher virion stability and increased in vitro SARS-CoV-2 replication fitness in primary human airway epithelial cells and Calu-3 cells [58]. Increased infectivity conferred by the D614G change is due to removal of hydrogen-bond interaction with T859 from a neighboring protomer of the S trimer. This results in an allosteric change of RDB domain to an "up" conformation that facilitates ACE2 receptor binding that may make the virus modestly more susceptible to neutralization by some sera or antibodies, depending the epitopes targeted by the antibodies [39,57,58]. Truncation of C-terminal 18 or 19 amino acids, which removes a putative ER retention signal, was also demonstrated to enhance HIV-based pseudotyping efficiency by 10-fold compared to WT S protein in 293T-ACE2 s cells [42,59]. The higher infectivity conferred by the C-terminal cytoplasmic tail truncation of 19 amino acids may be due to higher number of infectious particles [42].
Consistent with previous reports, we found that pseudoviruses bearing G614 and TR19 S proteins displayed 0.5-and 0.2-log higher infectivity, respectively, compared to WT pseudovirus in 293T-ACE2 s cells (Fig 1D). Pseudoviruses with TR14 S had slightly lower infectivity compared to WT pseudoviruses (Fig 1D). However, in 293T-ACE2.TMPRSS2 s cells the pseudovirus bearing the G614 S and TR19 S were more similar to WT S pseudoviruses, but the pseudovirus bearing the TR14 S displayed 0.5 log lower infectivity compared to WT pseudovirus (Fig 1D). Infectivity titers of all pseudoviruses were 1-1.5-log lower on 293T-ACE2 s cells compared to 293T-ACE2.TMPRSS2 s cells (Fig 1D). Based on these studies, we used WT S to qualify our pseudovirus neutralization assay because it represents the native, full-length spike, and the infectivity of WT S pseudoviruses give a large dynamic range for generating neutralization dose-response curves. Additional efforts to enhance pseudovirus infectivity with polybrene, a polycation that is known to enhance lentiviral transduction efficiency by minimizing charge-repulsion between the virus and cells, displayed no effect on pseudovirus infectivity.

Replication of SARS-CoV-2-mNG in 293T-ACE2.TMPRSS2 s cells
We confirmed the acceptability of the 293T-ACE2.TMPRSS2 s cells for SARS-CoV-2 infectivity and neutralization studies by assessing how well the cells support infection by replicating SARS-CoV-2. We compared replication levels and cytopathic effects (CPE) in 293T-ACE2. TMPRSS2 s to Vero E6 cells that are widely used for the propagation of SARS-CoV-2, as well as to 293T-ACE s cells that lack TMPRSS2. Typically, SARS-CoV-2-induces CPE in Vero cells by 48-72 h post infection (p.i.), characterized by cell rounding, detachment, degeneration, and syncytia [60]. However, by 24 h p.i., when both SARS-CoV-2-infected Vero E6 and 293T-ACE2 s cells began to display mNG expression and early CPE, SARS-CoV-2-infected 293T-ACE2.TMPRSS2 s cells displayed robust mNG expression and higher levels of CPE with nearly 50% of infected monolayer undergoing detachment (Fig 2). The rapid kinetics of infection is consistent with a study indicating co-expression of ACE2 and TMPRSS2 synergistically increases SARS-CoV-2 or pseudovirus entry efficiency and another study showing that SARS--CoV-2-infected Vero E6/TMPRSS2 cells displayed rapid infection kinetics and higher viral yield compared to Vero E6 cells [40,61]. The comparatively slower kinetics of infection observed in SARS-CoV-2-infected Vero E6 and 293T-ACE2 s cells in the absence of TMPRSS2 suggests less efficient entry via the endosomal pathway [55].

Entry pathways of SARS-CoV-2 pseudoviruses
Next, we used chemical inhibitors to determine the protease that primes SARS-CoV-2 S protein for membrane fusion during pseudovirus entry in 293T-ACE2.TMPRSS2 s cells. SARS-CoV-2 can use pH-dependent and -independent pathways for cell entry [62]. In cells lacking TMPRSS2, SARS-CoV-2 relies on endosomal-pH-dependent cysteine protease, such as cathepsin L for S priming, while entry is predominantly dependent on priming by TMPRSS2 in natural airway cells, such as lung epithelial cells, type II pneumocytes [62,63]. We found that pseudovirus entry into 293T-ACE2 S cells, which lack TMPRSS2, was sensitive to the endosomal pH acidification inhibitor chloroquine, with a half maximal inhibitory concentration [IC 50 ] of 0.79μM, but relatively insensitive to a TMPRSS2 inhibitor camostat mesylate. In contrast, pseudovirus entry was sensitive to camostat mesylate [IC 50 : 0.88μM] in 293T-ACE2.TMPRSS2 s cells, but much less so to chloroquine treatment (Fig 3). Thus, SARS-CoV-2 predominantly uses TMPRSS2 for priming S during virus entry into the 293T-ACE2.TMPRSS2 s cells.

Optimization of SARS-CoV-2 pseudovirus inoculum for neutralization assays
We next determined the inoculum range that would assure consistent neutralization titers according to the law of mass action [64]. Serial dilutions of rabbit serum or mAb (10G6H5) were mixed with four different pseudovirus inoculums over a three-log range, prior to incubation with 293T-ACE2.TMPRSS2 S cells (Fig 4). Although 100% neutralization was achieved at high serum or mAb concentrations using a relatively low inoculum of 2 x 10 4 relative luciferase units/ml (RLU/ml), the dose-response curve displayed high variation at higher dilutions, precluding generation of a reliable curve for calculating 50% neutralization titers (Fig 4A and 4B). However, inoculums in the 4 x 10 5 -1.4 x 10 7 RLU/ml range generated overlapping curves with little variation (Fig 4C and 4D). These dose-response curves yielded 50% neutralization titers with serum inhibitory dilution (ID 50 ) and mAb inhibitory concentration (IC 50 ) values that varied less than 2-fold over all pseudovirus inoculums. Therefore, inoculums of 5 x 10 5 -1 x 10 7 RLU/ml were used for the neutralization assay.

Assessment of the influence of ACE and TMPRSS2 levels on neutralization titers
Because the 293T-ACE2.TMPRSS2 s cells may have higher levels of TMPRSS2, ACE2, or both, compared to some primary airway cells, we explored whether different levels of ACE2 and TMPRSS2 on target cells might influence neutralization titers. We transiently transfected 293T cells with ACE2 and TMPRSS2 to achieve low, medium, and high levels of ACE2 and TMPRSS2. Expression levels of ACE2 and TMPRSS2 were confirmed via flow cytometry ( Fig  5). Transfection of a higher plasmid concentration of ACE2 and TMPRSS2 resulted in an increase in the number of ACE2 + /TMPRSS2 + cells (Fig 5A) as well as cell surface expression (Fig 5B and 5C). Neutralization assays performed with rabbit sera against RBD or S1 subunit, murine mAb 10G6H5, as well as an NIBSC reference plasma (#20/130), showed no significant differences in neutralization titers among the target cells with different levels of ACE2 and TMPRSS2 (Table 1). Neutralization ID 50 titers of rabbit sera against RBD and S1 subunit ranged from 9377 to 10540 and 5462 to 6742, respectively. NIBSC reference plasma ID 50 titers ranged from 2355 to 3130, while negative control sera lacked neutralization activity. IC 50 values for the mAb 10G6H5 ranged from 0.119 to 0.197μg/ml. The 80% neutralization titers (ID 80 or IC 80 ) were also similar among target cells with different levels of ACE2 or TMPRSS2. These findings indicate that levels of ACE2 and TMPRSS2 may not have a significant impact on neutralization titers for many antibodies.

Assessment of neutralization specificity and range of antibody titers
We assessed assay specificity and range of antibody titers using sera with reported neutralization titers, as well as 15 plasma samples from patients hospitalized with COVID-19. Thirty sera collected before 2019 served as negative controls, along with a negative control reference plasma standard (NIBSC #20/126) (Fig 6). Positive controls included the focused concordance samples comprising four high, five medium, and five low neutralizing antibody titers. All negative control sera failed to neutralize SARS-CoV-2 pseudoviruses at the lowest dilution tested (1:40). Neutralization titers (ID 50 and ID 80 ) segregated into high, medium, and low groupings, consistent with reported titers (Fig 6A and 6B). Plasma for patients hospitalized with acute COVID-19 showed a wide range of titers, consistent with previous reports [65]. We note, however, that the presence of reverse transcriptase or integrase inhibitors in sera or plasma samples from persons on anti-retroviral therapy (ART) has the potential to interfere with lentiviral pseudovirus readout that is dependent on reverse transcription and integration of the reporter gene. We therefore use lentiviral pseudoviruses bearing an envelope protein from amphotropic MLV or VSV as an additional control when assessing clinical samples that could include subjects on therapeutic or preventive ART. Non-specific inhibition of MLV-or VSV-pseudoviruses identifies sera that cannot be evaluated using this assay.

Assessment of assay precision
To further qualify the assay performance, we assessed intermediate precision among three operators using blinded test samples that included the six NIBSC plasma standards, 15 sera samples from the focused concordance samples panel, and a blinded panel of 21 sera samples that was used in a survey to assess assay concordance among labs using various SARS-CoV-2 neutralization assays (https://dhvi.duke.edu/duke-team-implement-sars-cov-2-neutralizationassay-concordance-survey-laboratories-worldwide). Neutralization titers giving 50% or 80% inhibition (ID 50 and ID 80 , respectively) compared to control were used to calculate the %CV. Only positive samples (with at least 50% of titer results � 1:40) were included in the precision calculation, which excluded six samples for ID 50 and 13 samples for ID 80 .The overall %CV across all samples for ID 50 and ID 80 titers was 38.8% and 30.8%, respectively, when titers < 1:40 were imputed to be 1:20. We consider these results to be acceptable for a neutralization assay and adequate for most clinical studies. When titers <1:40 were excluded from the analysis, the %CV was 27.5% and 20.7%, respectively.

Assessment of inter-laboratory agreement
We used the samples with reported neutralization titers as a benchmark for assessing accuracy of our assay. Although the reported titers were generated using different neutralization assay formats, we nevertheless found a strong correlation between our titers and the neutralization titers reported for the focused concordance samples (Fig 7A and 7B) and several NIBSC reference standards (Fig 7C and 7D). These results provide assurance that our assay provides titers that correlate well with titers measured in other assay formats.

Conclusion
We describe optimized procedures and detailed performance characteristics of an HIV-based, lentiviral pseudovirus neutralization assay for SARS-CoV-2 using a stable 293T cell line expressing ACE2 and TMPRSS2. The assay is quantitative, has a large dynamic range, and generates titers that correlate well with titers generated in other assays. The safety and relative

PLOS ONE
SARS-CoV-2 lentiviral pseudovirus neutralization in 293T-ACE2.TMPRSS2 cells simplicity of the assay makes it a valuable and versatile tool for evaluating mAb potency and neutralizing antibody titers in a BSL-2 lab setting.